Developing general-purpose learning systems capable of learning and performing unseen tasks using multiple modalities including images, videos, and knowledge graphs. Focus on foundation models like GPT-4, ChatGPT, CLIP, and Flamingo.
Applying graph neural networks and advanced techniques to scientific domains including molecule language models, molecule optimization for drug discovery, weather forecasting, and modeling complex scientific data on non-Euclidean spaces.
Developing generative models for drug discovery, image generation, and video generation. Focus on Bayesian optimization and diffusion models for designing novel molecules and producing realistic images.
High-level computer vision for scene understanding, object recognition, human-object interaction detection, and visual question answering. Applications include smart cities, security, and anomaly detection.
3D deep learning based on point clouds, voxels, and polygonal meshes for robotics, autonomous driving, and scene reconstruction. Focus on implicit neural representations for representing complex geometric shapes.
Studying reliability and safety of deep neural networks through adversarial robustness, uncertainty quantification, and Bayesian approaches. Focus on guaranteeing safety in autonomous systems and physical world deployment.
Analyzing structured medical imaging data including DTI, HARDI, and EAPs for disease detection and progression modeling. Research includes few-shot detection, domain adaptation, and contrastive learning for limited medical data.
SugaFormer: Super-class guided Transformer for Zero-Shot Attribute Classification
LLaMo: Large Language Model-based Molecular Graph Assistant
VidChain: Chain-of-Tasks with Metric-based Direct Preference Optimization for Dense Video Captioning
DialogGSR: Generative Subgraph Retrieval for Knowledge Graph–Grounded Dialog Generation
NF-BO: Latent Bayesian Optimization via Autoregressive Normalizing Flows
InvBO: Inversion-based Latent Bayesian Optimization
Deformable Graph Transformer (DGT)
EfficientViM: Efficient Vision Mamba with Hidden State Mixer based State Space Duality
MCTF: Multi-criteria Token Fusion with One-step-ahead Attention for Efficient Vision
DAVI: Diffusion Prior-Based Amortized Variational Inference for Noisy Inverse Problems
UP-NeRF: Unconstrained Pose-Prior-Free Neural Radiance Fields
Semantic-Aware Implicit Template Learning via Part Deformation Consistency
DeepVideo-R1: Video Reinforcement Fine-Tuning via Difficulty-aware Regressive GRPO
출처: 연구실 홈페이지
현재 재학생
13명
최근 5년 졸업
0명
학위 과정 분포: 석박통합 13명 (대학원 13명)
대학원 13명 · 포닥·학부연구생 4명 별도
본 페이지는 연구실 규모 파악을 위한 집계 통계(구성원 수, 진로 카테고리, 학위 과정 분포)만 제공하며, 개별 학생의 이름·전적·취업처 등은 표시하지 않습니다. 학위 과정 분포는 모든 재학생의 과정이 명확히 분류된 경우에만 표시되며 (분류 미상 학생이 1명이라도 있으면 미표시), k≥5 익명성 조건을 충족할 때만 공개됩니다 (PIPA §58-2·§28-2 + 대법원 2014다235080).
본 연구는 다중 성분 광활성 구조를 이용하여 에너지 수확과 센싱 기능을 모두 갖춘 자가전력 이중기능 센서 장치를 개발합니다. 최적화된 장치는 실내 조건(LED 1000 lx)에서 경질 및 유연 유기태양전지의 출력 전력 밀도가 각각 81, 76 µW cm⁻²를 초과하며, 광전자 모드에서 130 dB 이상의 선형 동적 범위를 갖춘 단일 픽셀 이미지 센서로 실내 응용이 가능함을 입증합니다.
논문 데이터가 수집되면 연구 키워드가 자동 추출됩니다